Like father, like daughter. Both Chip Watson and his daughter Anne Watson love physics and computer science and don’t like having to choose between the two. They both love traveling to China to teach English to high school students. And they both chose Duke for postgraduate work in physics. Chip earned his PhD in 1980 working with Prof. Emeritus Bilpuch and Anne is beginning her graduate studies at Duke this fall.
Chip is the manager of the High-Performance Computing Group at Jefferson National Lab in Newport News, Virginia, where he has been since 1988. He decided to pursue a career in computing while he was a postdoc. “While I was at Duke doing my physics research I was also picking up a lot of useful skills in computing, working on data acquisition systems and analysis,” he says. “When I did a postdoc at Stony Brook I found the computing was more enjoyable and I engaged my fellow researchers more with my computing than with physics so I abandoned the publish-or-perish rat race and took on a supporting role in being a staff member working on computer systems. But I didn’t take those skills out into industry—I very much enjoy the physics research environment.”
The high-performance computers that Chip oversees are used to analyze experimental physics data and perform calculations for theorists who are studying quantum chromodynamics—how quarks and gluons interact—using a lattice approach to manage the incredible array of necessary calculations. At each point on a grid, a discrete equation represents the environment around that point. “Lattice QCD is a numerical approach to pulling predictions out of that fundamental theory,” Chip says. “It’s incredibly difficult and very demanding on the processors.” Right now, the theorists are simulating areas slightly larger than a nucleon, but they hope to soon be able to model a small nucleus.
In 2009, Chip directed a major overhaul: “I built a system big enough and powerful enough that if users would take the effort to rewrite some code they would get a enormous benefit,” he says. The lab used funds from a $5 million economic stimulus grant to purchase graphic processing units (GPUs), which were originally designed for video games. GPUs can handle all the calculations needed to keep track of thousands of objects in a video game and provide realistic images of those objects falling, bouncing off each other, spinning, and shattering. “It did require considerable rewriting of code to take advantage of these things,” he says. “It was a major effort, but it was too compelling to ignore.” He says the GPUs provide about 10 times the performance on the dollar compared to more conventional platforms.
“We have over 500 GPUs,” he says. “It’s not the biggest in the world, but it’s a pretty good size. Ours in unusual in that it’s dedicated to a particular science domain.”
As a child, Anne visited Jefferson Lab during open houses and she also went with her dad to work on sick days or teacher workdays. (Her mom is a professional violinist, and Anne also plays violin.) Her middle-school and high-school science fair projects were always related to either physics or computers—or both. She says, “My dad in the true science spirit would say, ‘I’ll help you set it up, but you’re doing all the experiment yourself.’” He also connected her with other mentors at Jefferson Lab.
In her last two years of high school, she attended a regional public school called New Horizons Regional Education Center that offered advanced science classes. “What really made me love physics was a great teacher there,” she says. After high school, she was a Park Scholar at NC State, where she double majored in physics and computer science.
While her dad ended up at Duke primarily because his advisor at Georgia Tech recommended it, Anne applied to and was accepted at a number of graduate programs.
“I ended up choosing Duke not just because of the feel of the place but because of the opportunities I have at Duke,” she says. “I’m not married to a particular subspecialty of physics and Duke has four or five different professors I would like to work with—they are very encouraging about exploring all those opportunities versus other schools that would only offer me the opportunity to do research in a particular kind of condensed matter. Another thing I like about Duke is that it’s interdisciplinary.” The Free Electron Laser was also a major draw, since she has some familiarity with the FEL at Jefferson Lab.
Before arriving in Durham, Anne spent one week at Oak Ridge National Lab in Tennessee, meeting with some of the physicists there and investigating whether there might be a project she could collaborate on during her time at Duke. “Working at one of the Department of Energy National Labs is something I could see myself doing for a career,” she says.
If Duke Physics runs in your family, let us know.
Mary-Russell Roberson is a freelance science writer who lives in Durham.